UE5

Man, most indie games created with any engine aren't successful. Also, this even more true in the mobile market where Unity dominates.
And this general rule, most book authors aren't wildly successful. Most of the movies are crap we didn't even hear about. There are thousands of movies produces every year, how many of them you watch? For how many you pay?

Plus, if we're talking about metrics of success with this or another engine, it's always nice to gather some data. It's hard in this particular case, I know. But if someone says "I don't think games made with engine X are successful", it's just an impression. Confirmation of biases. Or sometimes excuses not to learn another tech/engine/market. Which is extremely time-consuming, so our psyche would love to auto-confirm ourselves that we keep doing the right thing by sticking to the current tech/engine/market/studio.

The success of indie game studios is tightly connected to used technology. Can I make a game I want by using a given tech/engine? Does this tech help me and make me more productive?
Look at what you can create/produce with a given technology. Evaluate available technologies yourself, don't assume "it's not good for indie".

It doesn't matter which one you choose - most independent creators using giving technology don't turn a profit. Or they earn enough just to survive, find another investor or support them by doing freelance work for others.

 

That's not what I said, 'X engine games are not successful' I don't think you read all my posts above, it would take the biggest wall of text of them all to sum it up here.

You're really misunderstanding me because of that.
To quote one of my previous posts just above:

And i did gather data (mentioned above as well), made big lists on steam, spreadsheets, gone through thousands of commercially successful games on steam and checked what engine it uses over months.

You can make a successful game on unreal, you can make a successful game with unity, and you can make a successful game with Monogame/custom and it goes on and on

 

Oh, ok. I thought for a second you meant that... but... wasn't sure why to add this statement in the UE5 thread on Unity forums - thread entirely about comparing tech.

My bad, wrong assumptions!

 

UE threads on unity aren't just to talk about UE tech, it's the place to talk about how you feel about unity too. As confirmed by this thread and the previous one, and a deleted one
Lots of discussions to be had on these threads, good chance to share thoughts and learn

 

Yeah, I noticed that this community is much more "talkative" than Unreal forums (much more focused on specific topics).

 

Good to see the crowd is able to agree to disagree and so on

 

Thing is unless engine x or y directly impacted the success of a game in a way other engines couldn't, there’s almost no reason to bring it up. I haven't really come across any Unity/Unreal game (indie wise) where I thought their engine choice really mattered or somehow made the end result easier or better.

For the team who made Valheim, having A engine mattered, but I doubt a different choice would've changed their success drastically. Its not a story of working with the ideal tools at the ideal time, but rather just having solid ideas that have appeal. On the flipside, many indie games will flop – regardless of their engine.

Its like bringing up guitar brands. Yes, there is a difference, Fender make great guitars, but its not what sets Jimi Hendrix apart.

 

They kinda do impact directly. As the feature set is not equivalent. So making a wrong choice can prolong the development.

 

If it was that simple, we would have a solution since long for Unity. I propose you take a closer look to Metahuman before you oversimplify things...

 

Well that was an exceedingly interesting experiment!! My daughter having utilized one of her 3090 cards downloaded and installed UE5, I believe she is utilizing the other for crypto mining although I am not entirely sure what the use is.

After furiously clicking away at some tutorials or so I am told she came up with the following.

Her conclusion is one of excitement and both disappointment, as the framerates aren't exactly great. The other thing is she believes that blueprints looks very similar to the way playmaker and or bolt functions, so in theory, if you understand such tool it should be fairly easy to pick up. Now of course, transitioning into script or code would present some issue. I believe this is where 'verse' could come to the rescue, although this is quite beyond my expertise right now and in honestly is quite baffling. One of the pictures has a special post processing effect enabled. Hope everyone is having a spectacular weekend, it's sunny over here, and I am keen to break out my water based oils

 

Interesting but I think wildly different for everyone, depending on how much experience / time they put into learning it. For some furious clicks however, that's an impressive banana.


Anyone want to see how many physics-based bananas can be rendered in DOTS ?

 

I honestly have no idea, apparently it was a one click thing after she installed the bridge plugin. She then proceeded to watch a blueprint tutorial at 2x speed, and away she went, having time to look after the grandkids, and bake my second youngest a birthday cake!

 

This topic reminds me of that phrase that says: "The neighbor's grass is always greener"...

We want Unity to commit to resources equal to those of unreal, but in the forums of unreal, they want optimizations like Dots from Unity.

But in the meantime, in other engine forums, developers want the "engine" to be the same as UNITY AND UNREAL (I'm looking at you Godot, Lumberyard and CryEngine)

I think it is a great merit of Unity to always be compared to the unreal, as there are several other capable engines in the sector. and unity is always one that directly compares

In this and other topics about the unreal there is always a bit of anguish, but let's be realistic, I think that changing the engine just because the other engine launched brilliant things doesn't make sense, we are professionals, not amateurs. I also know that not everyone has time to learn another engine, but if you have time, it's worth a try. In the past, I tried to use Godot and unreal, but it never served me well. I've worked at Unity for 5 years and it works well for me, and I almost never had any problems with it, and when I had a problem, a quick google search, and the problem was solved ...

But I agree with some points raised here. I hope that unity can optimize and improve its workflow, finalize its SRP pipelines, improve the documentation.

 

Metahumans are not that great for creating characters at the moment. There aren't any clothes, no way to import your own stuff, and editor is awkward.

One NICE thing about the tech is that they have facial mocap software. However, that one is locked to Apple phones, which makes it fairly pointless for many people.

Regarding character creator, there used to be Fuse, there used to be Morph3D, but they were phased out.
https://forum.unity.com/threads/in-...er-system-mcs-male-and-female-avatars.385627/

So it is largely a matter of someone pouring enough cash into the development.

 

Meshes are not enabled nanite out of the box, u need to enable it. Saw plenty of ppl complaining about performance and when they actually enabled Nanite it turned magic on

 

There's some progress, we can't say that Unity doesn't have free dark UI theme anymore.

It does if a project would make a good use of them. Knowing both engines and their strengths it's very helpful. For ex. I've made a mistake when creating rhythm game in Unity that lacks any sort of non-realtime sound analysis. Wasted a lot of time reinventing the wheel when most of it was already in Unreal.

 

Been playing with this for the past few days, really amazing piece of tech.

But question still stands, will they ever be able to make it work with foliage and characters. If they can pull it off, that will be absolutely amazing, imagine importing film quality trees with actual geometry on leafs etc , or CGI level characters and they gonna work on consoles. Future looks nice

 

Good heavens, thank you so much that certainly did do the trick. I must confess watching the visualization of the noise on the meshes is quite hypnotic, almost like perusing the Mandelbrot set and it forever draws closer and closer.

Have a wonderful day!

 

I do recall, that foliage does work with 'nanite' turned off although I'm not entirely sure that is a solution and if this is feasible in the long run, it may even work turned on - we did not try yet? Apparently the thinness of the meshes are a problematic pain point for the illumination / instancing solution.

 

Epic has some magicians in their ranks, I am sure they will pull something 'Unreal' xD

 

this is exactly what i am keeping my eye on.

 

Nope, not equivalent - but you tend to have pros and cons on each side. Would be a short tread if it was so.

Point was that none of the games owe their success to the engine in use, so there’s little reason to look at it as a marker. Like someone winning a golf tournament using a shovel, wouldn’t make me reconsider changing irons.

Technical challenges differ between engines, but there’s S*** on both sides of the fence.

 

That is false.

Ditch the engines and develop the game from scratch. See how many games make it to release this way. And how much time it would take.

While there are pros and cons, a game definitely owes it success to the engine it is using, because without it chances of it never making it to the release would be significantly higher.

 

Already said them having A engine matters.

 

Lumen looks very nice, I think it fits in well overall with the importance Unreal places on good looking environments. I'm hopeful that they won't wait too long to give vegetation a boost also.
I don't recall the last time an engine got as much excitement from the community (on YouTube at least).
*Sigh* Looks like 4 cores and 16gb ain't gonna cut it anymore.

 

If Lumen is some kind of SSGI, working with grass would require some fake shadows and lighting to be fast since grass is small. HDRP 12 introduced first grass so there is future for HDRP feature complete.

 

Poking around in the source has left me with the impression that Lumen is a hybrid approach taking advantage of screen space, raytracing, and voxels. In fact in the Project Settings window there are options for configuring both the hardware raytracing for it as well as the software raytracing fallback.

https://github.com/EpicGames/Unreal.../Engine/Source/Runtime/Renderer/Private/Lumen

 

Good news is that now that it's in early access there's documentation telling us how it works. Screen space stuff first, then voxelised distance fields.

Documentation is also available for Nanite, which is a hierarchical triangle mesh-based approach.

The other thing is that professionals see their off-the-shelf engine as a starting point, rather than a thing that should already do everything. Obviously it needs to give you the fundamentals to build on top of, otherwise it's not fit for purpose. I can't think of the last project I did where there the team didn't do a bunch of custom stuff on top of whatever tool set we were using.

Of course super efficient static meshes and improved lighting are both things that would certainly make an improvement to the fundamentals. They address very common use cases for many projects. The approach has been published now so, as with many other game dev technologies, I'm sure we'll start to see other implementations pretty soon.

 

So I'm more or less still new to this whole game dev thing, though I've done 3d stuff for at least over a decade.

Personally fancy visuals are whatever to me. You can have the latest greatest high fidelity visuals, but they'll still look like bunk in a few years. Yes people are into that, and I suppose that's where a lot of money is, but I'd much rather pursue paths of style and aesthetic that will last the tests of time. Kind of why my 3d style lately has gravitated towards mode toon/cell shaded (alongside it being a lot quicker to produce solo).

I do agree with the sentiments about Unity's development tools. It's weird to a new user like me to find everything has to be downloaded separately from the package manager, all in various states of development with separate documentation each time.

I've not used Unreal, not since it was called UDK or whatnot some 10 years ago. Though I'd be curious to give these demos a whirl on my beefy if old hardware to see how it runs.

 

One of the major advantages of Nanite is that it can scale up with future hardware and resolution advancements whereas current methods of implementing LODs are locked to hardware and resolution targets determined at the time you're designing the game.

A great example of a game that would have benefitted tremendously is Crysis. If Crysis would have had access to Nanite it would have looked every bit as good while having good performance too. Instead because it was based on traditional LODs it performed very poorly for its visuals.

Neither Nanite nor Lumen require that you be focused on photorealism. Unreal's demos are photorealistic because that's what attracts the most attention but if they wanted to they could have made just about anything utilizing the technology.

 

Unity has a lot of great talents. I follow some of unity graphics engineer on twitter. One guy make SDF renderer which allow million of SDF models render on the scene and each model has million of point clouds converted from million polygons. It's written in rust not C# and not running on unity. Maybe it's personal project or proof of concept? I'm not mention their name due to their privacy. Just said that unity has a great talent engineer. Maybe they're not in the position to decide.

 

Over the years, I've started to come to a similar conclusion. While our current project is PBR mixed in with stylized assets, I'm very eager to go full NPR (non photorealistic rendering) for our next project. Aesthetics are far more important than realism and offer better longevity as well. In the end it's all about the impression, and you don't need realism for that. As a matter of fact, it can get in the way.

What's so intriguing about UE5 for me isn't necessarily the realistic rendering aspects of it (although I can see that being useful for a ton of projects), but rather the improved workflow for developers in general. Not having to bake lighting is already a huge deal, but not having to care about poly counts on top of that is just icing on the cake. It makes the lives of developers so much easier, while at the same time making me question my work in Unity. Why do I have to spend time placing these probes? Why do I have to wait X amount of time for bake? Why do I have to spend time optimizing this asset? It all removes valuable time I could have used for creation rather than technical bs.

 

Ofc Unity has to have good talent too, You need to be talented and gifted engineer to produce a game engine, since its extremly complicated piece of software. I think the problem is, their talent may not be utilized to the full potential they have.

 

In this case what you you mean by "like Nanite"?

Do keep in mind that when you chose Unity circumstances were quite different, and that placing probes and doing bakes and all that other fun stuff was the standard way of approaching these things in many engines for quite some time - including in Unreal. Being disappointed by old stuff when comparing it to new shiny things isn't really being fair on yourself.

Also note that the new stuff is so new that it's not even ready yet, it's just in early access. It also comes with pretty hefty limitations: no Xbox One, no PS4, presumably no Switch. So in reality many games are going to have to keep doing things the current way for quite some time, as a) the new way matures and b) hardware which can use it becomes common enough to target exclusively.

Also, while I'm sure they're not as technically fancy, there are existing GI solutions which are bake-free and do work on current, common hardware, such as this one. When Unity announced they were doing their own thing I assumed it would be some variation on that.

 

Yes, that is true. It's not quite fair to compare the two, considering the circumstances were different at the time when I made this decision. However, the point I was trying to make was more along the lines of "the work I'm doing here is actually unnecessary in another tool", which makes pouring time into that particular kind work even more insufferable than it used to be.

I sure there are plenty of ways Unity could achieve better lighting and GI. I just don't know how much more hoping my heart can take. So many years and so little to show for it. It feels like I'm sitting at the edge of my seat waiting for an announcement that never came.

 

Crysis was designed to showcase an engine and thus the LOD targets were way ahead of what hardware of the time was able to properly handle. Nanite has the advantage that it is capable of generating LOD targets for both current and future hardware so if they had a similar like it the tech demo wouldn't have been quite as heavy.

 

This is specifically what I mean, though. It's no more unnecessary today than it was back when you made the decision. Even if you started your project today you'd still have to consider if you were willing to ignore current gen console users, which for many indie projects would be a clear "no".

Oh, definitely don't wait for that kind of thing. If it's not there to evaluate then don't rely on it coming later. Either do it yourself, find a 3rd party solution, or pick another toolset where it's already available. That kind of thing is inherently risky.

To be fair Crytek thought their stuff was targeting future hardware, which just happened to take a different direction.

 

Well that's actually quite neat to note. I'll totally admit I haven't looked too deep into this, though the general gist from what I saw it was all fancy, shiny, photorealistic stuff.

That said, thinking on it more, toon/cell shaded stuff does benefit from smooth high poly surfaces. So all in all, this could definitely be a benefit for non-photorealistic rendering styles. Now I'm intrigued.

 

Good heavens! This is totally relatable. I feel exactly the same way about 'art.' Having been classically trained, with knowledge of Renaissance and baroque techniques, when I began it was a much different affair.

Our skilled teacher at the time told us about the importance of observing the spaces in-between the objects when drawing and we spent an eternity with Pencils and shading along with understanding forms.

Colors came at a much later date, in fact, I believe there is much to be achieved when adopting a proclivity for restraint rather than un-restraint - if that made any sense.

As students, if can master the premises of the old masters you will observe that the old masters simply had an adept understanding of the fundamentals - that was it.

Toon shading is a wonderful idea, I suppose, popular mainstream shows, like Howl's moving castle and spirited away have fueled such passions, but my daughter mentioned unity has a open source project, called 'chop chop' from which you may derive certain ideas. That being said, of course, there are no limitations for unreal engine to achieve such a look, their shader graph looks very much alike to unity's one.

 

What wonderful little find, thank you so much for this!

 

During the last stream, they said that the Nanite demo works well even with HDD, so the previous consoles would handle it too. Probably with a lower resolution of screen/assets, as every 3D game.

Switch and mobile are a challenge, but they're working on the proper fallback. So we could run a UE5-powered game on every platform. Again, it's might be enough to decrease the resolution of assets and compress during cooking for a specific platform.

That's what Epic is repeating like a mantra. You not gonna be locked to current consoles if you'd use UE5 and Nanite. After all, Epic is making billions of dollars on making that one game on all possible platforms. And only $100 million/year on licensing Unreal itself

Can't find a specific timestamp... It's 2h long presentation + Q&A. It's worth of time if someone is interested how UE5 and what's comes beyond Early Access release.

 

We're talking about different things, though. That little tangent was about baking, probes, etc, and of Lumen the docs I linked earlier specifically say:

It also says that Nanite works on "PlayStation 5, Xbox Series S|X, and PCs with graphics cards meeting these specifications...". I haven't checked if PS4/Xbox One meet the specs, but it doesn't really matter since pushing out the mesh data in another format during the build process is certainly something which could be automated.

 

The docs state nanite proxy meshes are useful for collision and machines not capable of doing nanite such as mobile so it's all there right now. The docs also state they are improving the workflow for that so it can merge instances etc. So your mobile game still uses nanite to make, but classic meshes to render.

So it'll be fine, just not for a year, which is their deadline to polish that up. Personally, I don't think working top to bottom results in better looking games on low end hardware. I think working bottom up will be nicer obviously as the art is designed for the machine's capabilities.

 

And Lumen is designed to be able to fallback to their Screen space GI, Lightmapping. Obviously if your game heavily relies on dynamic GI, Lightmapping is out of the question, but maybe don't target mobile / low end then.

In any case, I like how there are much higher degrees of scrutiny for Unreal features in early access than for Unity features that were branded production ready years ago and are still dumpster fires.

 

That really sounds like even with Nanite you're going to need a rebuild of the game for the future generation.

Basically, the issue here is if you have a high detail model, the detail has to be stored somewhere. If the maximum level of details being used is lower than the level of detail stored, then the game is wasting storage space for something that is never utilized.

Which means that even with Nanite you'll reach the point where additional detail can't be provided, because there's no data on the disk for that.

So, what you'll end up is a title that is, say, made for 4k, looks okay in 8k, but then (that's ab exaggeration) 128k will come along and the title will look old-school.

Of course in reality the most likely sceanrio is that there will never be a 128k, and the game resolution and fidelity will stall around some point we're going to reach fairly soon.

 

Everyone is lumping on Unity here but with good reason. However, I downloaded the latest Unity 2021.1.9f1 (64-bit) (what a mouthful and 9f1s??) and chose an URP new scene. All gravy so far. Lovely experience with all the trash hidden from sight on package manager.

Things are slowly becoming hippo approved again. But there has been a lot of frankly piss poor decisions made at Unity and I've had enough, so future things Unity does must above all prioritise a stable front with hard-to-reach instable experimental areas. This will address a lot of concerns.

But I don't think Unity's learned their lessons yet. They still do things like make authoritative judgements on xyz feature alternatives or polls that make absolutely no sense, by people freshly hired without a clue. I hate that. It wastes my valuable time and drags down my will to help out.

What Unity needs to do is have stronger, clearer AAA management. Not an experimental employee in preview.

 

I don't have to. I have been doing this long enough to know my trade inside out. Face topology is face topology, a topology that currently holds 7 billion variants in the real world. Get a good face rig controlling that and some proper shaders, a library of major morph targets with new face rig deformer positions. toss a set of sliders to control percentiles of morph targets to blend in and you have the face variants. Hire me and I will build it for you. I did this already in C4D circa 2004 and could change from a little old oriental woman to an arnie big jaw neanderthal browed northern european to a lithe american female teenager. I could extract the morph at any point and feed it back through the pipeline to get more variants.

 

Yep, these always look like internships to me. And rarely hearing from them back. Like no more active on the forum.
I even wanted to make comment about that, but I restrained myself.

 

Don't restrain yourself. Unity staff even told me to comment more. Everyone at Unity is powerless to change unless the community is much more vocal. This is why defenders have no worth to Unity. So long as it's constructive, it's good.

 

Spoiler: rant, read at your own risk

To be honest I'm sick of seeing hardcore PR blogs from unity, sometimes It's literally everything I see in the blog.
(obviously not to say what they're talking about isn't important, or a priority) and while I understand the people writing it are most likely/most of them not developers, It gets annoying. Management is too busy saying the world is better with more creators in it and repeating it like a mantra and shoving the world's problems up your face to make it very clear that they care, they're different.

You don't see that amount of feel-good PR in unreal's blog, godot's blog, or cryengine
(Nowhere near as much anyways).
They're too busy focusing on the engine and talking about it, not some bullshit PR.

Epic's too busy giving a 100m to developers (I don't care where they got it from), acquiring products and making them cheaper/free for unreal users, and cheaper for everyone else too.
actually helping them -- not bullshit, overdone, seen it a thousand times "we're human too" "we care, to the extent we'll post about you, sort of, in the blog for human points" PR.
I checked the blog and there isn't as many as sometime back, but as usual there are still some posts.

Unity acquires products, and, if you're lucky it doesn't get any worse -- don't even think of it becoming cheaper or free or adding support/integration to unity users.

I have never in my life met a developer (or anyone for that matter) that said anything good about unity that was related to that PR work, as if your reputation will improve.

I don't know why it bothers me, but it just feels like such empty behaviour to make a corporation look caring from the human side of things.